Increased ICAM-1 Expression Causes Endothelial Cell Leakiness, Cytoskeletal Reorganization and Junctional Alterations  Paul R. Clark, Thomas D. Manes, Jordan S. Pober, Martin S. Kluger  Journal of Investigative Dermatology  Volume 127, Issue 4, Pages 762-774 (April 2007) DOI: 10.1038/sj.jid.5700670 Copyright © 2007 The Society for Investigative Dermatology, Inc Terms and Conditions

Figure 1 Adhesion molecule expression and permeability measurements on HDMEC transductants. (a) Adhesion molecule expression as assessed by direct immunofluorescence and FACS analysis is presented in histograms plotting fluorescence intensity (logarithmic x axis) versus cell number (y axis). Immunostaining with anti-ICAM-1 and anti-E-sel (dark lines), isotype-matched control (light lines) and the mean fluorescence intensity (MFI) corrected for isotype control are shown. (b) TEER properties of HDMEC transductants. TEER measurements on ICAMmed (top) and ICAMhigh (bottom) HDMEC transductants were recorded daily and compared to those on ICAMbasal (empty vector control transduced) and E-sel HDMEC. TEER levels (y axis) are the mean electrical resistance in ohms±SEM corrected for the resistance of a naked transwell. *P<0.001 by two-way ANOVA, Bonferroni post-test (n=6, 6, 6). Shown are two out of a total of five independent experiments involving three separately derived HDMEC lines for each construct. (c) Comparison of TEER with flux of BSA across transwells of transduced HDMEC monolayers. By FACS analysis HDMEC expressed ICAM-1 at basal, medium, or high levels. After measuring TEER, FITC-tagged BSA was added to the upper chamber of each transwell. The molar amount of BSA collected from the bottom chamber was used to determine a flux rate based on the transwell surface area (0.3cm2) and the duration of the flux assay (20hours). *P<0.001 by one-way ANOVA, Bonferroni post-test (n=6, 6, 6). Representative of two independent experiments. Journal of Investigative Dermatology 2007 127, 762-774DOI: (10.1038/sj.jid.5700670) Copyright © 2007 The Society for Investigative Dermatology, Inc Terms and Conditions

Figure 2 Cytoskeletal re-organization among HDMEC transductants. (a) Changes in actin cytoskeleton organization among HDMEC transductants were visually assessed by phalloidin staining of filaments and 4′,6-diamidino-2-phenylindole, dihydrochloride staining of nuclei. The actin cytoskeleton organization of ICAMbasal, ICAMhigh, ICAMmed, and E-sel HDMEC are shown. Bar=10μm. Representative of three independent experiments using two separately derived HDMEC lines. (b) Immunoblot analysis of cytoskeleton re-organization. The bands shown represent the relative amounts of F- and monomeric G- forms of β-actin, derived from detergent insoluble (insol.) and detergent soluble (sol.) protein lysates, respectively (see Materials and Methods), obtained from ICAM-1 and E-sel transductants (left) and non-transduced HDMEC either treated with vehicle or TNF (10ng/ml for 24hours; right). The F-/G-actin ratio shown below was determined by densitometry for each HDMEC condition after detection with mouse anti-β-actin mAb. Representative of three or more independent experiments with three different lines of transduced HDMEC and four different lines of non-transduced HDMEC. Journal of Investigative Dermatology 2007 127, 762-774DOI: (10.1038/sj.jid.5700670) Copyright © 2007 The Society for Investigative Dermatology, Inc Terms and Conditions

Figure 3 Assessment of changes in adherens and tight junctions among HDMEC transductants. The adherens junctions and tight junctions of HDMEC transductants fixed in ethanol/5% acetic acid were visualized by immunostaining with goat anti-VE-cadherin and mouse anti-ZO-1, respectively. Fragmented and interdigitized junctional immunostaining of VE-cadherin and ZO-1 in ICAMhigh differed from that of E-sel (arrows) which appeared continuous, more like that of non-transduced ICAMbasal control HDMEC. Antibody specificity was assessed by pre-adsorbing anti-VE-cadherin with immunogenic peptide before immunostaining (top left inset) and by comparing anti-ZO-1 with a mouse IgG1 isotype control (bottom left inset). Bar=10μm. Journal of Investigative Dermatology 2007 127, 762-774DOI: (10.1038/sj.jid.5700670) Copyright © 2007 The Society for Investigative Dermatology, Inc Terms and Conditions

Figure 4 TEER properties of ICAMΔCyto HDMEC. TEER measurements were recorded daily on ICAMΔCyto HDMEC and compared to those on vector-transduced ICAM-1basal, E-sel, and ICAMhigh HDMEC. TEER levels (y axis) are the mean electrical resistance in ohms±SEM corrected for the resistance of a naked transwell. *P<0.01 versus ICAMbasal by two-way ANOVA, Bonferroni post-test, n=5 per transductant. Representative of a total of six independent experiments involving two differently derived HDMEC ICAMΔCyto transductants. Journal of Investigative Dermatology 2007 127, 762-774DOI: (10.1038/sj.jid.5700670) Copyright © 2007 The Society for Investigative Dermatology, Inc Terms and Conditions

Figure 5 HDMEC responses to TNF. (a–c) Time dependence of HDMEC TNF responses. (a) FACS analysis of ICAM-1 expression through 24hours of TNF quantified as the isotype-corrected MFI of anti-ICAM-1-FITC immunostaining (logarithmic y axis). (b) TEER measurements of TNF-induced leakiness. TEER was reduced at 4hours of TNF (10ng/ml) versus at 10minutes of thrombin (2 NIHU/ml). The cytokine-induced fall in TEER is expressed as the percentage different from TEER of vehicle-treated control HDMEC (percent change, y axis; described in Materials and Methods).**P<0.001 for TNF (n=6) versus control (n=5), *P<0.001 for Thrombin (n=6) versus control by two-way ANOVA, Bonferroni post-test. Representative of three independent experiments with similar results. (c) Quantitative morphometry of TNF-induced shape change. TNF-induced shape change was delayed, first detected after 4hours of TNF treatment. Replicate monolayers of untreated HDMEC showed no change in shape over the same period (data not shown). *P<0.01 versus baseline (pre-TNF treatment); NS, not significant by one-way ANOVA, Bonferroni post-test. Representative of three independent experiments. (d–f) Concentration dependence of TNF responses. (d) By FACS analysis, ICAM-1 expression increased as a function of TNF concentration, reaching 50% of the measured change at 1.1ng/ml TNF. (e) TEER was reduced by 50% at 0.13ng/ml TNF, an approximately 10-fold less concentration than that required for a 50% increase in ICAM-1 expression or cell shape. The TNF-induced fall in TEER was calculated relative to the TEER of vehicle-treated HDMEC as in Figure 5b, except is expressed as a positive value for comparison with the HDMEC ICAM-1 and shape responses to TNF (percent reduction, y axis) (f) By morphometry, HDMEC shape elongated 50% at a TNF concentration of 1.1ng/ml. Data in (d–f) are representative of two independent experiments with HDMEC obtained from different skin sources. Journal of Investigative Dermatology 2007 127, 762-774DOI: (10.1038/sj.jid.5700670) Copyright © 2007 The Society for Investigative Dermatology, Inc Terms and Conditions

Figure 6 Responses of SRIκB- and enhanced green fluorescent protein-transduced HDMEC to TNF. TNF-induced leakiness is blocked in SRIκB HDMEC. Leakiness in response to either 10ng/ml TNF (top) or 2 NIHU/ml thrombin (bottom) is expressed as a percent of TEER measured on vehicle-treated replicate cultures (y axis) *P<0.01 by two-way ANOVA, Bonferroni post-test. Representative of three independent experiments and HDMEC obtained from two different skin sources. Journal of Investigative Dermatology 2007 127, 762-774DOI: (10.1038/sj.jid.5700670) Copyright © 2007 The Society for Investigative Dermatology, Inc Terms and Conditions